Potent Thromboxane A2 Antagonists - American Chemical Society

18-26 g, purchased from Hilltop Laboratories were utilized. Animals were randomly selected and allowed to acclimate to the laboratory environment for ...
0 downloads 0 Views 2MB Size
974

J . Med. Chem. 1989,32, 974-984

N-(2-Tetrazol-2-ylethyl)-4-[ (carbomethoxyoxy)methyl]4-(N-propionanilido)piperidine(4f): eluant EtOAc. NMR: 6 8.53 (s, 1 H), 7.37 (br s, 5 H), 4.98 (s, 2 H), 4.81-4.55 (m, 2 H), 3.84 (s, 3 H), 3.14-1.28 (complex, 12 H), 0.85 (t, 3 H, J = 7.0 Hz). N-(Phthalamidoethy1)-4-[ (propionyloxy)methyl]-4-(Npropionani1ido)piperidine (5a): eluant EtOAc-hexane 3:2. NMR: 6 7.90 (br s, 4 H), 7.42 (br s, 5 H), 4.98 (br s, 2 H), 3.71 (t, 3 H, J = 6.0 Hz), 2.85-1.62 (complex, 14 H), 1.47-0.92 (m, 6 HI. Pharmacological Methods. Mouse Hot-Plate Determination of ED, Dose. Ten male Swiss-Webster mice, weighing 18-26 g, purchased from Hilltop Laboratories were utilized. Animals were randomly selected and allowed to acclimate to the laboratory environment for at least 1 h prior to testing. Drugs were injected iv in the lateral tail vein. The animal was placed on a hot plate kept a t constant temperature (55.0 f 0.5 "C) and observed for occurrence of licking of hind or front paws. A timer was simultaneously started as the animal was placed on the hot plate and stopped when a nociceptive response was elicited. The animal was immediately removed from the hot plate and the latency time was recorded. Mice control latency times in excess of 15 s were eliminated from the study. Test latency times were determined 5 min after iv administration of the test drug. Analgesia is defined as test hot plate latency of 2 or more times greater than control latency times. No animal was permitted to remain on the hot plate longer than 30 s to prevent tissue damage. The number of mice exhibiting analgesia out of 10 was plotted as a percentage affected, thus generating a dose-response curve. At least three doses between 10 and 90% of responding animals were utilized. ED, values with 95% confidence limits were calculated." D u r a t i o n of Analgesia. Two times the ED,, dose was administered to 10 mice, and the hot-plate latencies were determined at various times after injection of the drug into the lateral tail vein. The mean maximum percent effect (% MPE) was calculated for each time period, and a time effect curve was generated. We have defined a test compound to be short acting if the duration

to 50% MPE was less than 6 min, intermediate with a duration of 6.1-15 min, and long acting with a duration greater than 15.1 min. Loss of Righting12 (LOR). Male Swiss-Webstermice weighing 18-26 g with free access to food and water were marked, weighed, and allowed to acclimate for 1h in the laboratory environment. The starting dose was usually 15.5 mg/kg with use of three mice per dose. The range of doses tested was that dose that elicits death in 1 / 3 mice (lowest lethal dose LLD) decreasing to a dose that does not produce loss of righting (LOR) in 0 / 3 mice. Immediately following a bolus injection into one of the lateral tail veins, a stopwatch was started and the mouse was placed on his dorsal side. Failure to return to the ventral side indicates LOR. Duration of LOR was recorded as the time from LOR until righting occurred. The lowest dose required to produce loss of righting in 3/3 mice (lowest effective dose 100, LED,,) was used as a standard for comparing the test compounds. Opiate Receptor Binding. The method was based on that of PasternakI3 et al. and used crude membrane fractions prepared from freshly harvested rat brains. Doses of 1, 10, and 100 nM were run in triplicate, and the data curve was fitted to the mean. The K ifor inhibition of [3H]nalaxonebinding was then calculated according to the method of Cheng and P r ~ s 0 f f . l ~

Acknowledgment. The technical assistance of M a r k Benvenga, Thomas Jerrussi, S t e v e Waters, and B a r r y I. Gold (Anaquest Pharmacology Department, M u r r a y Hill, NJ) and of t h e Dental School of the University of M a r y land at Baltimore is gratefully acknowledged. We also thank the discussions and valuable insight of Dr. Jerome R. Bagley (Anaquest C h e m i s t r y D e p a r t m e n t ) d u r i n g the course of t h i s investigation. Supplementary Material Available: A listing of the NMR data for 1-5 (4 pages). Ordering information is given on any current masthead page. (12) Method developed by F. G . Rudo, University of Maryland at

(11) (a) Tallarida, R. S.; Murray, R. B. Manual of Pharmacologic

Calculations with Computer Programs;Springer-Verlag: New York, 1984. (b) Domer, H. R. Animal Experiments in Pharmacological Analysis; Charles C. Thomas: Springfield, 1971; p 283.

Baltimore, Baltimore, MD 21201. (13) Ling, G . S. F.; Speigel, K.; Nishimura, S. L.; Pasternak, G. W. Eur. J . Pharmacol. 1983,86, 487. (14) Cheng, Y. C.; Prusoff, W. W. Biochem. Pharmacol. 1973,55, 11.

9,l l-Epoxy-9-homo-14-thiaprost-5-enoic Acid Derivatives: Potent Thromboxane A2 Antagonists S t e v e n E. Hall,* W e n - C h i n g Han, Don N. Harris, A n d e r s H e d b e r g , and M a r t i n L. Ogletree Departments of Chemistry and Pharmacology, Squibb Institute for Medical Research, P.O. Box 4000, Princeton, New Jersey 08543-4000. Received December 21, 1987 A novel bicyclic prostaglandin analogue, (1s)[ la,2a(Z),3a,4a]-7- [3-[ (hexylthio)methyl]-7-oxabicyclo[2.2. I ] hept2-yl1-5-heptenoicacid ((-)-lo), and its cogeners were found to be potent antagonists at the TxAz receptor. Compound (-)-lo was the only stereoisomer out of eight possible structures that was active. Thioether (-)-lo was 30-40-fold more potent than another TxAz antagonist, BM 13.177, in inhibiting arachidonic acid (AA) induced aggregation of human platelet-rich plasma. Compound (-)-IO was effective (Zm = 0.5 f 0.4 pM) in inhibiting 9,1l-amPGH2-induced (0.1 pg/mL) contraction of guinea pig tracheal spirals. The bronchoconstriction in anesthetized guinea pigs induced by AA was also effectively antagonized by (-)-lo (1 mg/kg, iv); however, in this assay (-)-lo exhibited some direct agonist activity. Radioligand binding studies in washed (human) platelets revealed that (-)-IO is one of the most potent ligands for the PGH2/TxA2 receptor yet described (Kd = 1.6 f 0.4 nM).

T h e development of pharmacological agents that modu l a t e t h e synthesis or actions of a variety of arachidonic acid (AA) metabolites c o n t i n u e s t o b e a n active a r e a of 0022-2623/89/1832-0974$01.50/0

research. O u r interest in t h e AA manifold has focused on t h e least s t a b l e m e m b e r of t h i s family, n a m e l y thromboxane A, (TxA2, l ) . I T h i s c o m p o u n d , whose s t r u c t u r a l C 1989 American Chemical Society

Journal of Medicinal Chemistry, 1989, Vol. 32, No. 5 975

Potent Thromboxane A , Antagonists

Scheme I

de

synthase q

:

OH x

A

2 a

2 (PGH,)

o d q O

a

1 (TxA2)

\

3

PGI, synthase

rfqH 6H

R.

H

3

\

Scheme 11"

h

4

4

6

endo-H

endo-H

1

endo-H

exo-H

8

exo-H

9

exo-H

endo-H

2

d,c,f

I]

d.e.f

4

BZ

10

endo-H

endo-H

11

endo-H

exo-H

12

exo-H

exo-H

13

exo-H

endo-H

* (a) TsC1, Py, CH2C12,23 "C; (b) KOtBu, RSH, THF, A; (c) 1 N LiOH, H20, THF, 23 "C; (d) PCC, NaOAc, CH2C12,Celite; (e) cat. NaOCH,, CH,OH, 0 23 "C; (f) NaBH4, CH30H, 0 "C.

-

OH

Scheme 111" 6H 4

3

l b

5

x=o x=s

assignment has been supported by total synthesis,2 is a potent stimulator of blood platelet aggregation and is also a potent smooth muscle spasmogen. Proof for a causative relationship between endogenous production of TxA2 and the pathophysiology of disease remains elusive due in part to the lack of specific tools that block the pharmacological effects of these endogenous mediators. A number of research groups have investigated the development of TxA2 synthetase inhibitor^.^ These agents block the conversion of the endoperoxide PGH2 (2) to TxA2 (Scheme I). Proponents of this approach have assumed that the PGH2 produced on stimulation would diffuse out of the cell responsible for its production (i.e. platelets) and be converted to the antiaggregatory and vasodilatory prostacyclin (PG12) by other cell types (i.e. e n d ~ t h e l i u m ) . ~A major disadvantage of this approach is that PGH2 possesses a similar pharmacological profile to that of TxA2 and is only 10-fold less potent2 than TxA2. This may be responsible, in part, for the disappointing clinical trials with these agents.5 An

w

q

c

7

0

r

b,c

wc0 16

6

14

-

R = CHI

IS R = C 4 H 9

(a) DIAD, Ph3P, HSAc, THF, 0 23 "C; (b) TsC1, Py, CH2C12; (c) KSAc, THF, DMSO, A; (d) KOH, C1CH2SR,xylene, & (e) 1 N LiOH, H20, THF. Scheme IV'

e.f,g 21

(1) Hamberg, M.; Svensson, J.; Samuelsson, B. Proc. Natl. Acad. Sci. U.S.A. 1975, 72, 2994. (2) (a) Bhagwat, S. S.; Hamann, P. R.; Still, W. C. J. Am. Chem. SOC.1985,107,6372. (b) Bhagwat, S. S.; Hamann, P. R.; Still, W. C.; Bunting, S.; Fitzpatrick, F. A. Nature (London) 1985, 315, 511. (3) (a) Iizuka, K.; Akahane, K.; Momose, D.; Nakazawa, M.; Ta-

nouchi, T.; Kawamura, M.; Ohyama, I.; Kajiwara, I.; Iguchi, Y.; Okada, T.; Taniguchi, K.; Miyamoto, T.; Hayashi, M. J. Med. Chem. 1981,24, 1139. (b) Tanouchi, T.; Kawamura, M.; Ohyama, 1.; Kajiwara, I.; Iguchi, Y.; Okada, T.; Miyamoto, T.; Taniguchi, K.; Hayashi, M.; Iizuka, K.; Nakazawa, M. J . Med. Chem. 1981, 24, 1149. (c) Gorman, R. R.; Johnson, R. A.; Spilman, C. H.; Aiken, J. W. Prostaglandins 1983,26,325. (d) Ford, N. F.; Browne, L. J.; Campbell, T.; Gemenden, C.; Goldstein, R.; Gude, C.; Wasley, J. W. F. J . Med. Chem. 1985, 28, 164. (e) Cross, P. E.; Dickinson, R. P.; Parry, M. J.; Randall, M. J. J . Med. Chem. 1986,29, 1637. (f) Manley, P. W.; Allanson, N. M.; Booth, R. F. G.; Buckle, P. E.; Kuzniar, E. J.; Lad, N.; Lai, S. M. F.; Lunt, D. 0.;Tuffin, D. P. J. Med. Chem. 1987,30,1588. (g) Press, J. B.; Wright, W. B., Jr.; Chan, P. S.; Haug, M. F.; Marsico, J. W.; Tomcufcik, A. S. J.Med. Chem. 1987,30,1036. (h) Akahane, K.; Momose, D.; Iizuka, K.; Mujamoto, T.; Hayashi, M.; Iwase, K.; Moriguchi, I. Eur. J. Med. Chem.-Chim. Ther. 1984, 19, 85. (i) For a recent review of TxA2 synthetase inhibitors and T d 2receptor antagonists; see: Cross, P. E.; Dickinson, R. P. Annu. Rep. Med. Chem. 1987, 22, 95. (4) (a) Gorman, R. R.; Bundy, G. L.; Pederson, D. C.; Sun, F. F.;

Miller, 0. V.; Fitzpatrick, F. A. Proc. Natl. Acad. Sci. U.S.A. 1977, 74,4007. (b) Needleman, P. Nature (London) 1979,279, 14. (c) Aiken, J. W.; Shebuski, R. J.; Miller, 0. V.; Gorman, R. R. J . Pharmacol. Exp. Ther. 1981,219,299. (d) FitzGerald, G. A.; Reilly, I. A,; Pederson, A. K. Circulation 1985, 72, 1194.

(a) PCC, NaOAc, Celite; (b) Ph3P+CH20CH3Br-,KOtAm, THF; (c) 20% TFA, THF; (d) NaBH4, CH30H, CeCl,; (e) TsC1, Py, CH2C12; (f) KOtBu, C5H11SH, THF, A; (g) 1 N LiOH, H20, THF. a

alternate approach is the pursuit of compounds that would block the actions of TxA, at the receptor level. This type of compound would also be expected to antagonize the effects of PGH2 since these two agents appear to share a common receptor.6 Sprague' and Nakane8 from these ( 5 ) (a) Belch, J. J. F.; Cormie, J.; Newman, P.; McClaren, M.;

Barbenel, L. J.; Capell, H.; Lieberman, F. P.; Forbes, C.; Prentice, C. R. M. Br. J . Clin. Pharm. 1983, 15, 1135. (b) Luderer, J. R.; Nicholas, G. G.; Nonamier, M. M.; Riley, D. L.; Vary, J. E.; Garcia, G.; Schneck, D. W. Clin. Pharmacol. Ther. 1984, 36, 105. (c) Reilly, I. A. G.; Doran, J.; Smith, B.; FitzGerald, G. A. Clin. Res. 1985, 33, 287A. (d) Zipser, R. D.; Kronborg, I.; Rector, W.; Reynolds, T.; Daskalopoulis, G. Gastroenterology 1984,87,1228. (e) McGibney, D.; Menys, V. C.; Nelson, G. I. C.; Kumar, E. B.; Taylor, S. H.; Davies, J. A. Clin. Sci. 1983, 12. (f) Reuben, S. R.; Kuan, P.; Cairns, T.; Gysle, 0. H. Br. J. Clin.Pharmacol. 1983,15, 835. (9) Kiff, P. S.; Bergman, G.; Atkinson, L.; Jewitt, D. E.; Westwick, J.; Kakkar, V. W. Br. J . Clin. Pharmacol. 1983, 15, 783. (h) Hendra, T.; Collins, P.; Penny, W.; Sheridan, D. Lancet 1983, 1041. (6) (a) Fitzgerald, D. J.; Doran, J.; Jackson, E.; FitzGerald, G. A. J. Clin. Invest. 1986, 77, 496. (b) Friedhoff, L. T.; Manning, J.; Cooper, W.; Willard, D. A. Thrombos. Haemostas 1985,54, 269. (c)LeBreton, G. C.; Venton, D. L.; Enke, S. E.; Halushka, P. V. Proc. Natl. Acad. Sci. U.S.A. 1979, 76, 4097.

976 Journal of Medicinal Chemistry, 1989, Vol. 32, No. 5

Hall et al.

Table I. Synthesis and in Vitro Activity of Thia 7-0xabicyclo[2.2.l]heptylAcids

no. 10 21 24 25 31 32 33 34 35 36 37 38 39 (+)-lo (-)-lo

R

synthesis RSH, KOtBu, reaction yield, config equiv equiv time, h 7’0 at C(9) 3.0 1.1 5.5 84 R.S R;S g RS g R.S &? d R;S d RS d RS 3.0 1.1 7.0 94 R,S 1.1 4.5 90 R,S 3.0 3.0 1.1 6.5 98 R,S 1.1 4.5 3.0 84 R,S 2.6 1.2 4.0 79 R,S 1.2 3.75 2.5 99 R,S 3.4 1.2 2.5 81 R 3.6 1.3 2.5 78 S

40 41 42 43 44 15

5.0 5.0 3.0

45 46 47 48 49 50 14 51 52

2.0

Effect of Substitution 7.6 86 S 10.0 89 S 2.5 3.5 76 S 1.1 1.1

S 3.0

7.4 5.0 5.0 3.0 15.6 g

LO 5

2.2

7.5

74 72

s S

Effect of Chain Length 2.0 7.5 50 S 1.1 16 96 S 1.2 8.6 96 S 1.1 17 88 S 1.1 5 91 s 1.1 16.25 81 S 81 S 1.5 8 94 s 1.1 7.5 98 s

in vitro pharmacology” AAIPA:‘ contraction of rat mp, “C I,, pM stomach strip A,,* pM oil 3.8 39 oil 37 oil 35 oil 230 83-87 192 9% 39-43 63 21 56-58 25 8 47-50 20 oil 5.5 23 oil 5.0 0.2 0.1 oil 20 oil 86 0.7 41-43.5 547 31-35 83 6% 31-35 1.0 25 oil oil oil oil oil oil

67-68 49-50 oil oil oil oil oil oil 110-112

3.8 A , = 22 pM

10 19 72 0.7 38 5.4

30% 13%

0.7

10%

1.0 0.5 0.2 0.7 42 13

26% 32 28% 0.8

Effect of Unsaturation 3 1.1 6 73 s oil 53 1.3 26 6 65 S oil 1.1 1.1 46 54 1.1 8 55 60 S oil 5.7 4.3 1.1 5.75 89 S oil 0.2 56 0.2 10 86 S 1.1 57 0.04 31-33 0.7 For details of the methods used, see ref 17;none of the compounds was effective in inhibiting ADP (20pM)induced platelet aggregation (PA) with the exception of sulfides 42 and 43. *Concentration of test compound required to elicit 50% of the maximal contraction induced by 3 X M serotonin; when given as a percentage, this is the maximum contraction observed; n = 8 for all compounds tested. I , vs 800 pM arachidonic acid in human platelet-rich plasma (PRP); values represent single determinations. Prepared by NaI04 oxidation of (&)-lo; see the Experimental Section. eFMI; refers to fast-moving isomer as determined on silica gel TLC; 1, vs ADP (20pM) induced PA = 45 pM. fSMI; refers to slow-moving isomer as determined on silica gel TLC; IN vs ADP (20pM) induced PA = 59 pM. #Prepared as described in the Experimental Section.

laboratories have previously reported the synthesis of a series of 7-oxabicyclo[2.2.l]heptaneanalogue related to 3 that were found to be potent TxA2 antagonists. Recently, we described a series of 7-oxabicyclo[2.2.l]heptaneether analogues (4)that were found to possess potent inhibitory action on PG ~ y n t h e t a s e . ~As part of our continuing program to identify novel TxA2 antagonists, we now deSprague, P. W.; Heikes, J. E.; Gougoutas, J. Z.; Malley, M. F.; Harris, D. N.; Greenberg, R. J . Med. Chem. 1985,28, 1580. Nakane, M.; Reid, J.; Haslanger, M. F.; Garber, D. P.; Harris, D. N.; Ogletree, M. L.; Greenberg, R. In Aduances in Prostaglandin and Thromboxane Research; Samuelsson, B., Paoletti, R., Ramwell, P., Eds.; Raven: New York, 1985;Vol. 15,p 291. (a) Hall, S.E.; Han, W.-C.; Haslanger, M. F.; Harris, D. N.; Ogletree, M. L. J . Med. Chem. 1986,29,2335. (b) Harris, D. N.; Phillips, M. B.; Michel, I. M.; Goldenberg, H. J.; Steinbacher, T. E.; Ogletree, M. L.; Hall, s.E. Prostaglandins 1986, 31, 651.

scribe the development of a new class of TxA2 antagonists that employs a simple thioether moiety as the w-chain. Some of these compounds (5) were also shown to have modest activity in inhibiting TxA2 synthetase. Chemistry The preparation of the initial target thioethers proved to be straightforward given the availability of alcohol esters 6-9. These intermediates were prepared from the exo and endo Diels-Alder adducts of furan and maleic anhydride as described by Sprague et al.7 The majority of the thioethers were synthesized as outlined in Scheme 11. Tosylation of the requisite alcohol ester followed by displacement with potassium hexanethiolate and hydrolysis provided the target acids 10-13 in good yield. Subsequent synthesis of additional thioethers followed this three-step procedure with few exceptions (Table I). In general, the required mercaptans were not commercially available but

Potent Thromboxane A2 Antagonists

Journal of Medicinal Chemistry, 1989, Vol. 32, No. 5 977

Scheme V"

Scheme VII. Evaluation of Stereoisomers Y

19

22

R=H R = M J a

23 24

R=CH3 R=H

10 AAIPA 150

Y

3.8 &M

1 1 Iso = 115 KM

a (a) MsCl, Py, CH2C12;(b) KOtBu, C7H,,SH, DMSO/THF, A; (c) 1 N LiOH, H20, THF.

Scheme VIa

26

21

+Y==--'L 0/

OH

0

25 er30 29

Rl = C,H,s. R2 = H RI - C 7H1sv Rz = CH3, C,H,s R I = Ac. R2 = CH,

-

" (a) NaH, Ph3PtN(CH3)PhI-, PhSH; (b) NaH, Ph3P+N(CH3)PhI-, C7H,,SH; (c) DIAD, Ph3P, HSAc, THF, 0 23 "C, 12 h; (d) KOH, xylene, C7HI5Br,A; (e) 1 N LiOH, HzO, THF.

-

-

were prepared by the method of Volante'O (ROH RSAc RSH) and were employed in the displacement step without purification. The target compounds that contained an additional heteroatom /3 to the sulfur at position 14 (i.e. 14 and 15) were necessarily prepared by an alternative route (Scheme 111). The Mitsunobu style conversion of alcohol 6 to thioacetate 16 proceeded efficiently (90%) but purification proved to be troublesome. Thus, a two-step process was used (R-OH R-OTs R-SAC) to provide 16 in 74% overall yield. Attempts to hydrolyze the thioacetate were plagued with formation of the cyclic sulfide 17 as well as the desired mercaptan. Synthesis of 14 and 15 was realized, however, by subjecting thioacetate 16 to the alkylation conditions that we found to work well in the 0-alkylationg of 6-9. In this reaction, the liberated thiolate is alkylated before it has the opportunity to participate in unwanted side reactions. A priori, it was not obvious what effect translocation of the sulfur atom might have with regard to TxAz antagonism. As such, we were interested in analogues of 10 wherein the sulfur atom at position 14 had been shifted to either the 13- or 15-position. A t the outset, this appeared to be a simple task as we had recently synthesized the prerequisite alcohols 18, 19, and 20.9 Indeed, conversion of alcohol 18 to thioether 21 was accomplished without incident under the conditions described above for alcohols 6-9 (Scheme IV). Preparation of the epimeric 13-thia analogues was, as expected, more difficult due to the increased steric congestion of alcohols 19 (Scheme Y) and 20 (Scheme VI). Displacement of mesylate 22 with heptyl mercaptan using the same conditions employed for the 14-thia analogues afforded only traces of 23. Addition of DMSO led to a modest yield (26%) of the thioether along with 38% of recovered mesylate 22 and 26% of the precursor alcohol 19. The NMR spectrum of 23 confirmed that the relative stereochemistry of the two side chains was

-

-

(10) Volante, R. P. Tetrahedron Lett. 1981, 22, 3119.

not scrambled as both H(9) and H(11-) appeared as doublets (coupled only to H(l0aa)). Had the trans isomer been formed, one of these resonances'(H-11) would appear as a triplet (also coupled to H(12)). No further work on this reaction was performed as there was enough product for initial pharmacological evaluation. Synthesis of the trans isomer 25 proceeded as shown in Scheme VI. Murahashi et al." had used a double-activation procedure to prepare the endo phenyl thioether 26 directly from the exo alcohol 27. In our hands, treatment of 20 under these conditions resulted in extensive degradation with lactone 28 obtained as the only isolable product (10% yield). Introduction of the sulfur was achieved on treatment of 20 with excess Ph,P/HSAc/diisopropyl azodicarboxylate.1° Thioacetate 29 was obtained in 33% yield along with 56% of recovered 20. Direct conversion of 29 to the desired thioether 30 was accomplished by treatment of 29 with KOH/heptyl bromide in refluxing xylene. Basic hydrolysis of the resulting mixture of methyl and heptyl esters afforded the target acid 25. Pharmacology In Vitro. Initial efforts to prepare analogues of either PGHz or TxAz have focused on alteration of the nucleus while maintaining the natural CY- and w-chains. Nearly every combination of nitrogen, oxygen, sulfur, and carbon atoms that would lead to a stable [2.2.1] or [3.1.1] bicyclic framework has been synthesized and evaluated. In general, analogues that possessed the natural a- and w-chains maintained proaggregatory/vasoconstrictor activity.12 In contrast, analogues that possessed a saturated a-chain13 or a modified w-chain (mainly replacement of the allylic alcohol with a diene or simple olefin) led to compounds that possessed activity as either TxAz synthetase inhibitors14J" or TxAzantag~nists.'~Many of these compounds (11) Tanigawa, Y.; Kanamaru, H.; Murahashi, S. I. Tetrahedron Lett. 1975, 16, 4655. (12) (a) Corey, E. J.; Nicolaou, K. C.; Machida, Y.; Malmsten, C. L.; Samuelsson, B. Proc. Natl. Acad. Sci. U.S.A. 1975, 72, 3355. (b) Corey, E. J.; Narasaka, K.; Shibasaki, M. J . Am. Chem. SOC. 1976,98, 6417. (c) Corey, E. J.; Shibasaki, M.; Nicolaou,

K. C.; Malmsten, C. L.; Samuelsson, B. Tetrahedron Lett. 1976, 737. (d) Miyake, H.; Iguchi, S.; Itoh, H.; Hiyashi, M. J. Am. Chem. SOC. 1977,99,3536. (e) Portoghese, P. S.; Larson, D. L.; Abatjoglou, A. G.; Dunham, E. W.; Gerrad, J. M.; White, J. G. J. Med. Chem. 1977,20,320. (f) Bundy, G. L. Tetrahedron Lett. 1975, 1957. (13) Raz, A.; Minkes, M. S.; Needleman, P. Biochem. Biophys. Acta 1977, 488, 305. (14) (a) Barraclough, P. Tetrahedron Lett. 1980, 1897. (b) Bundy, G. L.; Peterson, D. C. Tetrahedron Lett. 1978, 41. (15) (a) Kam, S.-T.;Portoghese, P. S.; Gerrad, J. M.; Dunham, E. W. J. Med. Chem. 1979,22,1402. (b) Gorman, R. R.; Bundy, G. L.; Peterson, D. C.; Sun, F. F.; Miller, 0. V.; Fitzpatrick, F. A. Proc. Natl. Acad. Sci. U.S.A. 1977, 74, 4007. (c) Shimizu, K.; Kohli, J. D.; Goldberg, L. I.; Kittisopikul, S.; Freid, J. In Advances in Prostaglandin and Thromboxane Research; Samuelsson, B., Paoletti, R., Ramwell, P., Eds.; Raven: New York, 1983; Vol. 11, p 333. (d) Maxey, K. M.; Bundy, G. L. Tetrahedron Lett. 1980, 445.

978 Journal of Medicinal Chemistry, 1989, Vol. 32, No. 5

were only evaluated for their activity in the platelet; however, some of these analogues (Le. carbo-TxA,) possessed potent vasoconstrictor activity despite their antiaggregatory activity in the platelet.I6 The compounds in the present study are similar to these analgoues in that the allylic alcohol has been replaced by a more lipophilic residue, namely, an alkyl sulfide. All of the thioethers were evaluated for their ability to inhibit both arachidonic acid induced and adenosine diphosphate induced platelet aggregation (AAIPA and ADPIPA, respkctively) of human platelet-rich plasma (PRP)." These results are summarized in Table I. Initial evaluation of the four isomeric. thioethers 10-13 established that only the cis-exo isomer possessed significant activity (Scheme VII). The poor activity of the trans-isomer 1 1 was surprising since the analogues prepared in the above studies possessed the same relative stere~chemistry'~-'~ as sulfide 11. Translocation of the sulfur atom from position 14 to either 13 or 15 led to a 10-fold decrease in activity. As a result, our initial structure-activity relationships (SAR) were developed with a series of 14-thia cis-exo analogues (racemic). Oxidation at sulfur was not tolerated as evidenced by decreases in activity ranging from 10- to 50-fold for sulfoxides 31 and 32 as well as the sulfone 33. Likewise, modification of the thioether alkyl residue led to dramatic attenuations in activity. For example, homologation of the alkyl group in the w-chain from six to eight atoms (10 vs 39) resulted in a 100-fold drop in potency. Evaluation of the individual enantiomers of 10 revealed that both antipodes possessed antiaggregatory activity. Further investigation revealed that (--)-lo (absolute stereochemistry as shown in Scheme VII) derived its activity as a thromboxane antagonist. Its antipode (+)-lo,in analogy to the related 0-ether analogue^,^ was found to inhibit prostaglandin synthase. To avoid possible complications in the interpretation of the results from racemic mixtures, all subsequent thioethers were prepared as single enantiomers. As our initial studies suggested that the antiaggregatory activity of the thioethers was strongly influenced by the nature of the sulfide w-chain, our SAR work focused on modification of the w-chain tail. These compounds are grouped in Table I into three general categories representative of the following changes: substitution in the alkyl chain, length of the alkyl chain, or unsaturation in the alkyl chain. Within the first category, substitution a or /3 to the sulfur atom led to decreased activity. The stereochemistry of substitution in the aposition (C-15) was important; whereas methylated analogue 40 was just slightly less potent than (-)-lo, the methyl epimer 41 was actually a stimulator of platelet aggregation.ls Unlike the other thioethers prepared, alcohols 42 and 43 possessed inhibitory activity against ADP-induced aggregation of human PRP. Within this (16) (a) Corey, E. J.; Ponder, J. W.; Ulrich, P. Tetrahedron Lett. 1980, 137. (b) Nicolaou, K. C.; Magolda, R. L.; Claremon, D. A. J . Am. Chem. SOC. 1980, 102, 1404. (c) Nicolaou, K. C.; Magolda, R. L.; Smith, J. B.; Aharony, A.; Smith, E. F.; Lefer, A. M. Proc. Natl. Acad. Sci. U.S.A. 1979, 76,2566. (d) Wilson, N. H.; Jones, R. L. In Aduances in Prostaglandin and Thromboxane Research; Samuelsson, B., Paoletti, R., Ramwell, P., Eds.; Raven: New York, 1985; Vol. 14, p 393. (e) Ansell, M. F.; Caton, M. P. C.; Palfreyman, M. N.; Stuttle, K. A. Tetrahedron Lett. 1979, 4497. (17) Arachidonic acid (800 pM) and ADP (20 fiM) induced platelet aggregation in platelet-rich plasma as described: Harris, D. N.; Phillips, M. B.; Michel, I. M.; Goldenberg, H. J.; Sprague, P. W.; Antonaccio, M. J. Prostaglandins 1981, 22, 295. (18) It is interesting to note that the antagonist/agonist profile observed for 40 and 41 parallels that found for allylic alcohol 3 (antagonist) and its carbinol epimer (agonist).

Hall et al. Table 11. Effect of Thioethers on TxB, and PGE, Synthesis in

a Lysed Platelet Preparationn drug concn, inhib of TxB, no. PM synthesis, 70 (-)-lo 1.0 7 3.1 3 10 17 31 29 100 69 (+)-lo 10 33 100 45 1000 75 21 10 33 100 80 1000 90 49 10 33

stimulation of PGE, synthesis, 90 174 137 230 404 556 27 44 -51* 13 -45b -75b 220 380 100 22.1 860 333 51.8 1650 1000 61.5 1820 56 10 210 33 420 100 24.4 900 333 56.9 1340 1000 91.2 1920 "Compounds 30-32 had no effect on either TxB, or PGE, synthesis. PGE, synthesis was inhibited; these compounds apparently inhibit cyclooxygenase.

limited group of compounds, the only modification tolerated was the isosteric replacement of a methylene group with an additional sulfur atom (15). Within a homologous series of alkyl thioethers some latitude in length was tolerated. Thioethers possessing alkyl chains of three to six carbons were equipotent, while alteration of the length beyond these limits led to sharp decreases in activity. In fact, the only modification that led to increased activity, albeit modest, was the introduction of unsaturation in the alkyl residue.lg The allylic thioethers (49,53-55)were similar or slightly more potent that their corresponding saturated analogues. The homoallylic thioethers (50,56,57)were approximately 5-fold more potent than their saturated counterparts. Given the effect of the related 0-ethersgon AA metabolism, several of the more potent thioethers were evaluated for their effects on AA metabolism in a lysed platelet preparation (Table 11). Some of these analogues possessed TxA2 synthetase inhibitory activity, but only at concentrations at least 10-fold higher than the concentration necessary for effective inhibition of platelet aggregation. This was clearly not due to cyclooxygenase inhibition since PGE, production was stimulated by over 500% at the concentrations that inhibited TxA, synthesis by 50% (60-310 pM). This result was not completely unexpected as a number of PGH2/TxAz analogues that possessed highly lipophilic w-chains inhibited TxA, synthetase.14JMJM The prospects of a pharmaceutical agent with both TxA, antagonist activity and TxA2 synthetase inhibitory activity are attractive as these compounds would not only block the effects of TxA, (as well as PGH2)at the receptor level, but also might shunt PGH, from the platelet to the vasculature resulting in the conversion of PGHz to PGIz. Critical to the development of a TxA2 receptor antagonist as a potential pharmaceutical agent is whether these agents possess any agonist activity. As a general screen for direct agonist activity, we evaluated the effects of the more potent thioethers on rat stomach strips.m This tissue (19) It was anticipated that introduction of a double bond in the a-chain might enhance the inhibitory activity of these thioethers toward TxA, synthetase, in analogy with the TxA, synthetase inhibitory activity exhibited by PGH,.

Potent Thromboxane A, Antagonists

Table 111. Antagonist Activity of Thioethers in Guinea Pig Tracheal Spirals"

U

150,FM histamine 9,11-azo-PGH2 >lo0 0.5 f 0.4 (-)-lo SCsHl3 >lo0 3.8 f 0.3 49 SCH&H=CH, >lo0 0.4 f 0.04b 53 (E)-SCH,CH=CHC3H-, >lo0 1.6 f O . l b 54 (E)-SCH,C(CH3)=CHC3H7 55 (Z)-SCHZCH=CHC3H7 >lo0 1.6 f 0.2 >lo0 0.38 f 0.02b 56 (Z)-SCH,CH,CH=CHCZH, "Contraction induced by either histamine (1 gg/mL) or 9,llazo-PGHz (0.1 gg/mL). For comparison, BM, 13,177,25a structurally unrelated TxAz antagonist, had no effect on histamine-induced contraction and was weakly active against 9,11-azo-PGHz (I5,,= 100 gM, 46% inhibition at 100 pM). bAntagonism of 9,llazo-PGH, responses remained after "washout". no.

R

is unique in that it contracts to prostanoids with few exceptions. Almost all of these thioethers displayed some direct contractile activity on the stomach strips. A t the outset it was assumed that the contractile activity was due to activation of the PGH2/TxAz receptor, but evidence to support this assumption remained to be identified (vide infra). The concentration-effect curves for the majority of the thioethers evaluated in the rat stomach were consistent with them functioning as partial agonists in this tissue since the magnitude of the contraction seldom exceeded 50% of the control response. Although the lack of a correlation between the TxA, antagonist activities of the thioethers, as measured in the platelet screen, and their contractile potencies in the rat stomach strip suggested that these two activities might be dissected, we were unable to identify analogues of (-)-lo that were completely free of contractile activity. Thus, (-)-lo and its congeners were shown to be qualitatively similar to a growing number of PGH,/TxA, analogues that, despite potent antagonist activity in the platelet, display varying amounts of PGH,/TxA2 agonist activity in a variety of smooth muscle preparations.16 As an alternative measure of TxA2 antagonist activity, several of the thioethers were evaluated for their ability to block contractions of guinea pig trachea spirals induced by either g,ll-azo-PGH, (a stable PGH2/TxA2 mimetic)12a~b or histamine (Table III).,' Consistent with their mechanism of action, all of the thioethers tested were effective in inhibiting 9,11-azo-PGH2-induced but not histamine-induced contractions. In general, the potencies of the thioethers in this assay correlated well with their activity in the platelet screen. Unlike the contractile activity displayed by these thioethers on rat stomach strips, none of the sulfides tested exhibited any direct effect on the guinea pig tracheal spirals. One interesting yet unexplained observation is that the inhibitory effect of several thioethers (54, 56) could not be readily washed from the trachea spirals. (20) Am determined as the concentration of test compound required

to elicit 50% of the maximal contraction of rat fundic stomach M serotonin. For details see: (a) strips induced by 3 X Ogletree, M. L.; Harris, D. N.; Greenberg, R.; Haslanger, M. F.; Nakane, M. J . Pharmcol. Exp. Ther. 1985,234,435. (b) Harris, D. N.; Greenberg, R.; Phillips, M. B.; Michel, I. M.; Goldenberg, H. J.; Haslanger, M. F.; Steinbacher, T. E. Eur. J . Pharmacol. 1984,103,9. (21) Contraction induced by 9,11-azo-PGH2 (0.1 gg/mL) or histamine (1.0 pg/mL) as described in ref 20.

Journal of Medicinal Chemistry, 1989, Vol. 32, No. 5 979 Table IV. One-Site Analysis of Inhibition of Specific Binding of HSQ in Washed Plateletsa Kd f SEM, nM slope factor f SEM no. (-)-lo 1.6 f 0.4 0.73 f 0.03 47 97 f 20 0.75 f 0.04 49 76 f 11 0.67 f 0.01 50 49 f 5 0.72 f 0.01 53 3.1 f 0.5 0.99 f 0.05 56 0.4 f 0.1 0.75 f 0.06 Inhibition of specific binding of HSQ (5,6-di-3H-SQ 29,548), [l~-[1~u,2ru(Z),3cu,4a]l-7-[3-[[2-[(phenylamino)carbonyl]hydrazino]acid as methyl]-7-oxabicyclo[2.2.l]hept-2-yl]-[5,6-3H~]-5-heptenoic described in ref 22b. Values reported are an average of four determinations. Table V. Evaluation of Antagonist/Agonist Activities of Thioethers

in Anesthetized Guinea Pigso inhibition of AA responses agonist activity, % duration, RL MABP RL, 5% MABP, % min no. (-)-lo +349 +lo7 95 153